PREDICTIVE CHEMINFORMATICS ANALYSIS OF DIVERSE CHEMOGENOMICS DATA SOURCES: APPLICATIONS TO DRUG DISCOVERY, ASSAY INTERFERENCE, AND TEXT MINING Stephen Joseph Capuzzi A dissertation submitted to the faculty at the University of North Carolina at Chapel Hill in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Division of Chemical Biology and Medicinal Chemistry in the Pharmaceutical Sciences Department in the Eshelman School of Pharmacy Chapel Hill 2018 Approved by: Alexander Tropsha Stephen V. Frye Albert A. Bowers Nikolay Dokholyan Dmitri Kireev © 2018 Stephen Joseph Capuzzi ALL RIGHTS RESERVED ii ABSTRACT Stephen Joseph Capuzzi: Predictive Cheminformatics Analysis of Diverse Chemogenomics Data Sources: Applications to drug discovery, assay interference, and text mining. (Under the direction of Alexander Tropsha) In this dissertation, we describe the cheminformatics analysis of diverse chemogenomics data sources as well as the application of these data to several drug discovery efforts. In Chapter 1, we describe the discovery and characterization of novel Ebola virus inhibitors through QSAR-based virtual screening. In Chapter 2, we report the discovery and analysis of a series of potent and selective doublecortin-like kinase 1 (DCLK1) inhibitors using QSAR modeling, virtual screening, Matched Molecular Pair Analysis (MMPA), and molecular docking. In Chapter 3, we performed a large-scale analysis of publicly available data in PubChem to probe the reliability and applicability of Pan-Assay INterference compoundS (PAINS) alerts, a popular computational drug screening tool. In Chapter 4, we explore the PubMed database as a novel source of biomedical data and describe the development of Chemotext, a publicly available web server capable of text-mining the published literature. iii To my mother. iv ACKNOWLEDGEMENTS I would like to thank my advisor, Dr. Alexander Tropsha. As Markovnikov wrote to Butlerov: “Считаю приличным посвятить небольшой труд свой Вам, многоуважаемый наставник, поскольку проводимые в нём мысли есть дальнейшее развитие установленного Вами….Если в нём и заключается что-нибудь новое, то рождение этого невозможно было бы без исходных положений, заложенных Вами.” I would also like to thank the various members of the Molecular Modeling Laboratory, especially Dr. Eugene Muratov. To my committee members, thank you all for your guidance, support and insights. I would also like to thank my friends and family – too many to name. Lastly, I would like to acknowledge my funding sources: NIH, NSF, DoD, USP, and EII. v TABLE OF CONTENTS ABSTRACT ................................................................................................................................................ iii LIST OF ABBREVIATIONS ................................................................................................................... xi CHAPTER 1: COMPUTER-AIDED DISCOVERY AND CHARACTERIZATION OF NOVEL EBOLA VIRUS INHIBITORS ................................................................................................................ 13 1.2 INTRODUCTION ......................................................................................................................... 13 1.3 RESULTS ..................................................................................................................................... 15 1.3.1 Model Performance ........................................................................................................... 15 1.3.2 QSAR-Based Virtual Screening ........................................................................................... 17 1.3.3 Experimental Evaluation .................................................................................................... 18 1.3.3.a Experimental confirmation of Anti-EBOV activity of 14 compounds .......................... 18 1.3.3.b Mechanisms of action against EBOV entry ................................................................. 21 1.3.4 Cheminformatics Analysis .................................................................................................. 25 1.3.4.a Assay Liabilities ........................................................................................................... 25 1.3.4.b Chemical Similarity to Training Set Compounds ......................................................... 26 1.3.4.c Comparison to Previously Reported EBOV Inhibitors ..................................................... 27 1.4 DISCUSSION ............................................................................................................................... 34 1.5 CONCLUSIONS ........................................................................................................................... 36 1.6 EXPERIMENTAL SECTION........................................................................................................... 37 1.6.1 Data Collection, Curation, and Classification ..................................................................... 37 vi 1.6.1.a Antiviral Data ............................................................................................................... 37 1.6.1.b Cytotoxicity Data ......................................................................................................... 38 1.6.1.c Determination of Antiviral Activity ............................................................................. 39 1.6.1.d Determination of Cytotoxicity .................................................................................... 40 1.6.1.e Antiviral Training Set Balancing .................................................................................. 40 1.6.1.f Cytotoxicity Training Set Balancing .............................................................................. 40 1.6.1.g Modelability Index (MODI) .......................................................................................... 41 1.6.2 Computational Methods .................................................................................................... 41 1.6.2.a QSAR Model Generation and Validation ..................................................................... 41 1.6.2.b Virtual Screening ......................................................................................................... 42 1.6.3 Experimental Methods ....................................................................................................... 43 1.6.3.a Materials ..................................................................................................................... 43 1.6.3.b Cell culture methods ................................................................................................... 51 1.6.3.c Ebola VLP beta-lactamase assay for HTS in 1536-well plates ..................................... 51 1.6.3.d Cell viability assay with the ATP content assay kit ...................................................... 52 1.6.3.e Ebola live virus assays ................................................................................................. 52 1.6.3.f Filipin staining and LysoTracker-red staining ............................................................... 52 1.6.3.g Cathepsin B/L assay ..................................................................................................... 53 1.6.3.h Thermal shift binding assay with Ebola VLP................................................................ 53 1.6.3.i Data analysis and statistics .......................................................................................... 53 vii 1.7 ASSOCIATED CONTENT.............................................................................................................. 54 CHAPTER 2: COMPUTATIONAL DISCOVERY AND EXPERIMENTAL VALIDATION OF POTENT INHIBITORS OF THE UNDERSTUDIED KINASE DCLK1 ........................................... 55 2.2 INTRODUCTION ......................................................................................................................... 55 2.3 RESULTS ..................................................................................................................................... 57 2.3.1 QSAR Model Development ................................................................................................ 57 2.3.2 QSAR-Based Virtual screening ........................................................................................... 59 2.3.3 Experimental Validation ..................................................................................................... 60 2.3.4 SAR Analysis and Implications for Future Design ............................................................... 62 2.3.4.a Matched Molecular Pair Analysis and Model Interpretation ..................................... 62 2.3.4.b Molecular Docking ...................................................................................................... 63 2.4 DISCUSSION ............................................................................................................................... 64 2.5 CONCLUSIONS ........................................................................................................................... 69 2.6 METHODS .................................................................................................................................. 70 2.6.1 Data Production, Collection, Curation, and Classification ................................................. 70 2.6.2 Computational Methods ...................................................................................................